are three particular problems with indicator-based approaches: the indicators are based on relationships (eg R&D inputs drive

Transcription

1 Innovation Systems in Australia 1 Don Scott-Kemmis, Innovation Management and Policy Program National Graduate School of Management, Australian National University 1. Innovation Systems Frameworks Countries differ in their capacity to produce, acquire and use knowledge. They differ in the level of their investment in innovation, the roles of the public and private sectors, the industries and technology fields of greatest importance and the rates of change in those patterns, the level of cooperation among organisations, the modes of financing innovation, attitudes to risk taking, the regulation of the labour market and the role of large and small firms. In short, they have different innovation systems. The structure, functioning and integration of the various components of the national innovation system (NIS) have a major bearing on the level, and continuing upgrading, of a nation s innovation competencies. These competencies play a central role in economic growth and change. Following the seminal works of Freeman (1987), Lundvall (1992) and Nelson (1993), there have been a large number of studies of national innovation systems. These studies have sought to analyse the ability of nations to generate, diffuse and use economically significant knowledge. However, there have not been any comprehensive studies that have analysed Australia from a national innovation systems perspective. How can we analyse Australia from an innovation systems perspective? The NIS approaches do not provide any ready to use frameworks. Studies that use a national innovation systems approach tend either to assume a high level of homogeneity within nations, or to focus on only some components of the NIS often the R&D system. Many studies use a range of innovation-related indicators. 2. This chapter outlines three complementary indicator-based analyses, each of which reveals important characteristics of the Australian innovation system: 1 This chapter is based on the AUSIS project which is supported by the Australian Research Council through the ARC Linkage Program. The partners on this project are: the Innovation Management and Policy Program, National Graduate School of Management, ANU, Department of Industry, Tourism and Resources, Department of Agriculture, Fisheries and Forestry, the Australian Business Foundation, the National Office of the Information Economy and the CSIRO. This paper is based on the work of the AUSIS project team in particular, Antonio Balaguer, Kevin Bryant Robert Dalitz, and Magnus Holmen. 2 While valuable in raising issues and questions to be investigated, indicator approaches rarely provide real insight and answers. There are three particular problems with indicator-based approaches: the indicators are based on relationships (eg R&D inputs drive ISRN 6 th Annual Conference Working Paper 1

2 A review of traditional innovation-based indicators, presented here in terms of two contrasting perspectives on Australian innovation performance; An analysis of patterns and trends in R&D activity at the national and state level; and An analysis of patterns and trends in Australia s technological and scientific specialization. The final section identifies several characteristics of the Australian innovation system, and then to develop several more speculative interpretations of the evidence. 2. Innovation Systems: An Analytical Perspective National innovation systems can be understood as a nation s capacity to generate, diffuse and use economically significant knowledge. Innovation systems evolve and a primary endogenous driving force is learning: if knowledge is the most important resource, then learning is the most important process (Lundvall, 1992). Learning is the central process not only in the generation of knowledge but also in the diffusion and use of knowledge. Learning occurs at the level of individuals and organizations but also through various forms of interactions between actors: eg firms, government agencies, universities, and formal or informal bridging organisations. Among the more important mechanisms for interactive learning are producer-user networks, the pooled labour market, informal meeting places, etc. These patterns of interaction draw on established networks based on antecedent innovations and production links that may only slowly include new actors. For this reason these interactions may lead to path-dependencies in innovative search. These learning processes are set by nation specific circumstances, including the accumulated skills and capabilities of firms, national laws and regulations, culture and the specialisation in research and education, and are derived from its history. Lundvall (1992) conceptualised a national innovation system as consisting of two main parts: the economic structure and the institutional set-up. The economic structure refers essentially to what a nation produces - products, services, technologies, the labour force, skills etc.. 3 The economic structure largely shapes what nations learn. For example, if a country has a large mining industry, then different actors are likely to learn a great deal about large innovation) that are themselves in question; aggregation leads to a serious loss of information; there are no indicators for many important categories. 3 Note that even if there is not a one-to-one relation, the economic structure reflects the technological and industrial specialisation of a country. To some extent then, the economic structure can be analysed through the means of taxonomies (Pavitt, 1984, Soete 1989, Miles and Coombs 1999) simply by associating data on the economic structure to the various categories in the taxonomies. ISRN 6 th Annual Conference Working Paper 2

3 projects and logistics. An economic structure may be more or less beneficial in terms of current market growth trends 4. The institutional set-up refers to the structure of organisations and institutions, and includes the nature and processes of (product, capital, labour, equity, IP) markets and networks, strategies of firms, the type of regulation and structure of incentives shaping, reinforcing or constraining the direction of the innovative search. 5 Hence, the institutional set-up refers to how the generation, diffusion and use of economically useful knowledge takes place. 3. The Australian Innovation System: Alternative views on performance Assessing Innovation Related Performance Indicators derive their meaning from assumptions about what phenomena are important and how they can best be estimated. The relationships between innovation and economic activity are complex, involve interactions, lags and feedbacks, and evolve continuously. Understanding these relationships requires attention to multiple dimensions, including institutional issues Deviations from the performance of a model economy (typically the USA) or from best practice exemplars among OECD countries in specific dimensions do not necessarily signal a problem. Countries are not arrayed along a path of inexorable development. Nevertheless, such comparisons (Figure 1 and Figure 2) can be useful for characterising innovation-related performance if we bear in mind the structure and history of the Australian economy: Diversity. Australia s population is concentrated in several cities distant from each other. Agriculture is a major industry but is diverse operating in temperate, tropical and semi-arid areas. Mining is a major industry, but across a continent mining is highly diverse involving coal, iron ore, gold, aluminium, silver, mineral sands, nickel, diamonds etc- each with different requirements for capital goods, services and infrastructure. Historical legacy. The legacy of import substitution industrialisation, was a range of institutions (eg labour market regulation, attitudes to entrepreneurship), innovation-related infrastructure (a well developed public education and research system, but poor researchindustry links in many sectors) and competences (low management and R&D capabilities in 4 For example, a high level of dependence on ICT exports was a good thing in 1999, but far less attractive in The concept organisation here covers all types of organisations, such as firms, policy agencies, universities etc. The notion institution refers to either pervasive or persistent patterns of behaviour in the nation or rules or regulations within a country. ISRN 6 th Annual Conference Working Paper 3

4 industry) that continued to impede innovation performance even when the incentives for innovation increased, following the opening of the economy 6. Industry structure. Primary products account for about 8% of Australian GDP (the level for about 30 years), above most OECD countries. The manufacturing share of GDP (about 12% in 2000) is lower and declining faster than in most OECD countries. The services sector is larger (79% of GDP in 2000 and has risen more rapidly than in most OECD countries. Australia has a relatively very small high-tech manufacturing sector. This sector, and particularly large firms in this sector, account for the majority of BERD, and most interaction with the public sector research system in most OECD countries. Firm size. Australian industry has a relatively large proportion of small firms and small firms (less than 100 employees) account for a relatively large share (almost 30% in 1999) of BERD. 7 Such small firms account for twice the proportion of BERD in Australia as in Canada or Finland, and three times more than in the US or the UK. Trade. Australia s trade intensity (trade/gdp) is relatively low, closer to the low trade intensity of large economies like the USA, Japan and France, than to the high trade intensity of small countries like Ireland, Finland, Sweden and Canada. Australia did not participate in the strong growth of trade in manufactures of the 1970s, losing opportunities to develop economies of scale. Specialisation. Australia has a relatively low level of technological specialisation for a small economy small advanced economies tend to be quite specialised in some fields of technology. Whereas most countries have become more specialised over the past 20 years, Australia s level of specialisation has remained more or less constant. Australian specialisation is in agriculture, primary metals, mining and oil & gas- a pattern quite similar to that of Canada and Norway. 8 Foreign ownership. The level of foreign investment in Australian industry is relatively high, and particularly high in the R&D intensive sectors. Overall, foreign affiliates account for almost half of the R&D in Australian manufacturing, a level far higher than all but a few other OECD countries 9 6 See for example: Elek, Camilleri & Lester. The Role of Technological Change in Australian Economic Performance. Centre for Economic Policy Research. ANU. 1987, and Sheehan, P. and G. Messinis. Innovation in Australia. 7 The Allen Consulting Group. Systematic Mismatches in the National Innovation System Aggregate patent performance is a poor indicator of Australian performance, because much technological activity is in fields of low patent activity and also activity in high tech fields tends to be in narrow niches rather than broad fields. 9 DITR Australian Innovation Scoreboard (2003) ISRN 6 th Annual Conference Working Paper 4

5 Figure 1. Australia's Innovation Performance Compared to OECD Average* Public R&D Scientific & technical articles per capita US patents Business R&D Labour force, % with tertiary education Number of science graduates aged Number of researchers in labour force Venture Capital as a % of GDP Number of Internet users Investment in new equipment as % of GDP Foreign affiliates in manufacturing R&D Strategic alliances Innovation as a % of total sales MFP growth Percentage difference from OECD Average Sources: ABS (2002a, 2002b, 2002c, 2002e), OECD (2002a, 2002d, 2002e, 2002f, 2002g, 2001a, 2001b, 2001c, 2001d, 2001e), US Patents & Trademark Office (2002) and World Competitiveness Yearbook (2002) Figure 2. Australia's Innovation Performance Compared to OECD Leader* Australia OECD Leader Ratio to OECD Average Public R&D Scientific & technical articles per capita US patents Business R&D Labour force, % with tertiary education Science graduates per 100,000 labour force Researchers in labour force Venture Capital % of GDP Internet users Investment in new equipment % GDP Foreign affiliates in manufacturing R&D Strategic alliances Innovation spending % total sales MFP growth Sources: ABS (2002a, 2002b, 2002c, 2002e), OECD (2002a, 2002d, 2002e, 2002f, 2002g, 2001a, 2001b, 2001c, 2001d, 2001e), US Patents & Trademark Office (2002) and World Competitiveness Yearbook (2002) It is possible to assess a wide range of indicators of the many aspects and dimensions of Australia s innovation system, and of innovation-related performance, and reach two quite different conclusions: either that the innovation system is robust and adaptive; or alternatively that it is weak and locked-in to old patterns of specialisation. ISRN 6 th Annual Conference Working Paper 5

6 Chart 1: Alternative Perspectives on Australia s Innovation Performance The Dynamic Growth Perspective Australia is a broadly based dynamic and flexible The Laggard Perspective The Australian economy maintains a high level of economy, diversified across markets, and increasingly sectors, underpinned by competitive domestic markets and flexible labour markets. High-level human resources and strong research organisations facilitate the rapid uptake of new knowledge produced anywhere. Imported knowledge and equipment combined with local knowledge and capability supports active problem solving and systems integration in a range of sectors generating relatively high levels of productivity. A fast-user strategy combined with natural and human resources is a sound basis for future prosperity. A focus on R&D and patents misses the level of dynamism in technology adaptation and application. Key indicators of this performance include: High and increasing productivity; Relatively high level of public sector R&D; Substantial growth in niches markets in key manufacturing sectors: telecom equipment, wine, boats, automobiles and components; Maintaining strong competitiveness in resources sectors through the effective application of new technology, including IT; Increasing technological specialisation in biotech and pharmaceuticals; High FDI as % of GDP; Strong performance in international science; A strong ICT services sector and high growth in knowledge based services ; Rapid and broadly-based uptake of ICT. dependence on natural resources and is failing to develop sustainable new areas of specialisation and growth. Productivity growth in the 90's is the result of one-offs: micro-economic reform and the uptake of ICT. This performance masks underlying weaknesses in new firm formation and in the growth of new internationally competitive industries. The poor performance of Australian firms in R&D and patenting signals the weaknesses in management, scale and international positioning of Australian industry. Australia s declining position in high tech sectors and the declining international significance of its science and patents indicates the extent to which Australia is being left behind the frontier of innovation and growth in the world economy. The 3 rd lowest ranked in the OECD in gross expenditure in R&D and one of the lowest in business R&D; The lowest expenditure on innovation among OECD countries; A relatively very low level of investment in venture capital; International patenting activity (per mill. pop n.) is one of the lowest in the OECD; 80% of the top fifteen export products are resource-based commodities with a low level of processing; A large and growing trade deficit in ICT products and services. ISRN 6 th Annual Conference Working Paper 6

7 Chart 3 Dynamic Growth Perspective Innovation-Related Investment Australia has been a leader in pubic investment in R&D and particularly in relative investments in basic research. Government investment in R&D as a proportion of GDP is slightly greater than the OECD average 10. Knowledge-based industries contribute 31% of GDP (in 2000) and knowledge workers represent 38% of the labour force levels similar to comparable OECD countries. The proportion of knowledge workers in the labour force increased at a similar rate to most other OECD countries 11. Expenditure on higher education (1.6% of GDP) is comparable with the OECD average.. 12 About 18% of the Australian workforce has tertiary qualifications, above the OECD average (14%) 13. About 24% of tertiary students in Australia are in science & technology fields, above Canada (16%) and the US, but below Finland (38%) and many other OECD countries. 14 Industry investment in workforce training increased strongly through the 1990s 15. Innovation-Related Performance Australian production of scientific and technical articles (about 700 per million population) is greater than the OECD average (about 450/million).16 MPF growth ( % pa over ), was comparable to or better than most OECD countries. 17 The growth in GDP per hour worked ( about 2% pa over ), was one of the highest among OECD countries. 18 By the early 1990s R&D intensity in some Australian manufacturing sectors (eg metal products, iron & steel, shipbuilding) was above the OECD average. The service sector in Australia was a particularly strong R&D performer 19. Linkages Foreign direct investment inflows over the 1990s (1.75% of GDP) were well above the OECD average (1.0%), but outflows (0.8% of GDP) were well below the OECD average (~ 1.4%). 20 Australia has Australia has been estimated to have one of the highest levels of international inter-firm alliances (about 5.5 per US$ billion GDP) in the OECD. This level is similar to that of Canada. 21 Over the past 20 years the rate of growth of Australia s trade intensity has been among the most rapid in the OECD, similar to Canada and Finland. The growth in export performance has been increasingly broadly-based in terms both of products and markets DEST Australian Science at a Glance DITR Australia as a Modern Economy, Estimates based on OECD categories. 12 Ibid p DITR Measuring Innovation Performance. Current Status and Future Considerations. Draft April DITR Australia as a Modern Economy, 2002; p36. Figures are for AEGIS The High Road and the Low Road. Australian Business Foundation DITR Australia as a Modern Economy, 2002; p45. Figures are for OECD STI Scoreboard, Ibid. 19 DIST Australian Business Innovation. AGPS DITR Australia as a Modern Economy, 2002; p20. Estimates based on OECD categories, for OECD STI Scoreboard, Figures are for the period. 22 Working Paper 4: Assessing Australia: Characteristics, Innovation, Structure and Specialisation ISRN 6 th Annual Conference Working Paper 7

8 Exploring New Innovation-Based Opportunities Australia is reported as having a high proportion of the population working in new firms (17%), significantly higher than Canada (11%) and Finland (9%). 23 Australia has the third highest level of expenditure on ICT (10.5% of GDP) in the OECD 24. The rise in ICT expenditure in the 90s has been comparable to other OECD countries. The share of ICT investment in total non-residential investment has risen steadily since the 1980s (22.5%) is the third highest in the OECD 25 The rate of growth over the 1990s of Australian biotech patenting in the US (about 18% pa) was one of the fastest in the OECD (av. 8% pa) 26 Australia has steadily increased its relative level of specialisation in most medical-related fields: pharmaceuticals, biotechnology, and medical instruments 27. Exports of knowledge based services have grown strongly over the 1990s, and as imports have declined as a proportion of GDP, net exports have grown more rapidly than for most other OECD countries 28. Australia s relative export specialisation in resource-based products (largely minimally processed) has increased over the past 30 years. Wine and boat building emerged in the 1990s as new areas of comparative specialisation Global Entrepreneurship Monitor Figures are for2001 and the data were developed from surveys. 24 DITR Australia as a Modern Economy, 2002, p3.. Estimates based on OECD categories, for DITR Australia as a Modern Economy, 2002, p5.. Estimates based on OECD categories, for US Patent and Trade Mark Office. Data for Working Paper 4: Assessing Australia: Characteristics, Innovation, Structure and Specialisation. 28 DITR Australia as a Modern Economy, Estimates based on OECD categories and cover the period , 29 Working Paper 4:Assessing Australia: Characteristics, Innovation, Structure and Specialisation. Based on Revealed Comparative Advantage scores. ISRN 6 th Annual Conference Working Paper 8

9 Chart 4 Laggard Perspective Innovation-Related Investment Business investment in R&D (about 0.65% of GDP) is one of the lowest in the OECD, less than half the OECD average (about 1.4%). Business R&D (BERD) grew strongly through the early 1990s, declined from 1995 to 2000 and has increased from 2000 to Overall R&D investment levels rose from the mid-1980s to the mid 1990s to levels (about 1.6% of GDP) significantly below the OECD average. Expenditure on innovation by manufacturing firms (estimated at about 1.9% of sales) was one of the lowest in the OECD. 31 A relatively low proportion of Australian managers hold tertiary qualifications 32. Innovation-Related Performance Australian patenting levels in the US (about 40 per million population) are comparatively very low, less than a third the level of Canada and Finland. Australian patenting in the US grew over the period at a rate similar to other OECD countries, but as a consequence the gap in patenting level widened. Australian patenting in the US is more widely dispersed, with fewer areas of high specialization than is the case for most other OECD countries. 33 Medium-high technology and particularly high technology industries account for a relatively small share of Australian exports, about 32% compared to the OECD average of about 65%. Despite a growth in trade intensity Australia s deficit in medium and particularly high tech products has widened through the 1990s 34. Over the period, employment growth in Australia was largely in sectors of low innovation-related investment (R&D, training) 35. Broadband penetration rates at 0.57 per 100 people are significantly below the OECD average (1.96 per 100 people) and comparable countries such as Canada (6.3%) 36. Linkages From the early 1990s the role of FDI inflows (%GDP) increased strongly in most OECD countries but not in Australia, where Australia s share of FDI inflows declined markedly. The relative size of the stock of FDI in Australia (FDI stock/gdp) is one of the highest in the OECD. 37 While Australia has a relatively high number of alliances, a relatively high proportion of these are domestic and a relatively low proportion are technological DITR Australia as a Modern Economy, 2002; p34-5. Figures are for Ibid; p36. Figures are for 1997 for Australia and 1996 for other countries. 32 Industry Task Force on Leadership and Management Skills. Enterprising Nation (Karpin Report) Canberra: AGPS; 1995) 33 Australian Research Council & CSIRO Inventing Our Future p27 and DITR Australia as a Modern Economy, 2002; p43 34 AEGIS (1997) The High Road and the Low Road. ABF. ;Sheehan, P. and G. Messinis. Innovation in Australia. 35 AEGIS op cit 36 DITR Australia as a Modern Economy, 2002; p13... Estimates based on OECD categories, for Ausis 2003 Working Paper 4:Assessing Australia: Characteristics, Innovation, Structure and Specialisation. Based on Revealed Comparative Advantage scores. IMPP. ANU 38 OECD (2001 and OECD STI Scoreboard (1999). The latter trend may be a result of the former and both may be due to Australia s relative isolation. ISRN 6 th Annual Conference Working Paper 9

10 Exploring New Innovation-Based Opportunities Investment in Venture Capital (about 0.06% of GDP) is below the OECD average (about 0.14% of GDP). 39 The proportion of venture capital directed to early stage funding appears to be relatively very low. Australian firms tend to focus less on innovation in products and services than do firms in other countries, and have markedly less confidence than do firms in other countries in capturing value from innovation. 40. Some evidence indicates that the level of entrepreneurial activity in Australia, while increasing, is lower than in many other OECD countries 41. While ICT imports (~3% of GDP) are the average level for the OECD, ICT export levels are relatively very low (1% of GDP) less than a third of the OECD average 42. Because Australia has a relatively very small ICT manufacturing sector, the share of ICT employment in business employment (about 4.6%) is one of the lowest in the OECD 43 The rate of growth over the 1990s of Australian ICT patenting in the US (about 10% pa) was below the OECD average ( 13% pa). 44 Australia has increased its relative level of patenting activity in the traditional resource-based fields 45 Overall Australian patenting tends to be in areas where technology is moving less rapidly and Australian patents tend to have a relatively high level of linkage to science but to be based on older prior knowledge 46 Australia has increased its relative export specialisation in resource-based products. Canada and Finland also have a comparative export specialisation in resource-based products (largely wood), but these are significantly processed prior to export and both of these countries have strongly increased their relative specialisation in high VA manufactured products 47. Characterising the Australian Innovation System. Before an initial assessment of the performance of Australia s NIS it is useful to draw out three systemic characteristics of the NIS: Resource-Enabled, Knowledge-Based, Competition-Driven Innovation. Mining and agricultural industries have a vital role in Australia s balance of trade. A substantial part of Australia s research system is linked to these industries, as are a wide range of manufacturing and service sector suppliers. The performance of much of Australia s mining and agricultural industries is dependent on innovation based on complex technologies and high-level 39 OECD STI Scoreboard, Figures are for the period. 40 David Brown.The Innovative Company. Arthur D. Little International OECD Economic Survey: Australia Paris OECD STI Scoreboard, Figures are for the period 43 DITR Australia as a Modern Economy, 2002; p7.. Estimates based on OECD categories, for US Patent and Trade Mark Office. Data for Ausis 2003 Working Paper 4: Assessing Australia: Characteristics, Innovation, Structure and Specialisation. IMPP. ANU 46 Ausis 2002 Working Paper 2: Innovation In Australia: Characterisation of Four Themes in Australian Innovation Systems.IMPP. ANU ISRN 6 th Annual Conference Working Paper 10

11 capabilities. These industries are resource-enabled but increasingly market and innovationdriven. In major areas of mining and agriculture Australian productivity performance is world leading. In both mining and agriculture the strong and sustained demands for innovation and problem solving have led to the emergence of specialist providers of equipment and services although much of the core capital goods are imported. Many of these specialist suppliers are now exporting goods and services. Dispersion, Fragmentation and Focussing Devices. Australia s innovation systems are highly dispersed: geographically, sectorally, technologically and organisationally. The scope for economies of scale in innovation and production has been more limited than the aggregate picture would suggest. The significance of barriers to focus, critical mass and effective interaction is generally underestimated. Systems Integration and Problem Solving in the Innovation System. A great deal of innovation in Australia involves essentially systems integration - combining subsystems and adapting systems to meet Australian needs. These processes often require highlevel capabilities to solve problems and incorporate novel design elements. 48 R&D and patent statistics tell us little about these types of innovation, which are central to productivity. As organisational change is often required for effective technological innovation, managerial will have a major bearing on the effectiveness of technological innovation and the returns to investment. Managerial competencies are also vital user-producer links and supply chain development which are increasingly associated with technological innovation. Three dimensions provide a useful starting point for assessing overall NIS performance: Performance in generating (and renovating) resources required by firms and other problemsolving organizations. These resources include human resources, knowledge, networks, infrastructure, trust, standards, etc. Performance in solving problems ie in mobilising resources to meet performance gaps. This operates at the level of the firm, the technological system, the sector and, in relation particularly to the policy domain, at the national level. Performance in ensuring diversity and hence generating options for economic (etc) progress ie building capacities beyond those needed for current problem solving, as in 47 Ausis 2003 Working Paper 4: Assessing Australia: Characteristics, Innovation, Structure and Specialisation. IMPP. ANU 48 With implications for the role of the public sector research system and approaches to its evaluation. ISRN 6 th Annual Conference Working Paper 11

12 developing new competencies, technological trajectories, industries, clusters, innovation systems 49.. In relation to all three of these dimensions it is clear that Australian innovation systems are evolving in response to the opening of domestic and international markets and technological trajectories. Over the 10 years from 1988 Australia moved from having some of the highest tariff levels in the OECD to generally the lowest levels 50. Substantial components of agriculture are shifting from commodity production by developing higher value-added activities based on differentiation in products and marketing-related services. Significant new areas of strength have developed in for example, wine, scientific and control instrumentation, and some services sectors. In relation to the first two of these dimensions the overall evidence suggests that the innovation system is performing reasonably well, when assessed in terms of recent performance. Australia has a strong public sector research and education system and is an effective user of new knowledge and technology from domestic and international sources 51. However, there are some critical caveats to make, including: growth off a low base (eg in some areas of patenting and trade); the drivers of export growth (including the role of the depreciation of the A$); and the sources of productivity growth (particularly the role of one off factors such as micro-economic reform), hence: It is beyond dispute that Australia has experienced strong economic growth for a decade or more, during the latter part of which time many standard indicators of innovation have been falling. Again these outcomes can be interpreted in different ways. One view is that these facts show that broad economic change can be more important to growth than innovation as commonly measured, because market forces will find the best growth opportunities. The other view is that Australia s recent growth spurt is unsustainable, being driven by rapid growth in borrowings by households, by a surge in net foreign debt and by a falling dollar.. On this account, these and other factors have masked Australia s declining position in the global knowledge economy. (Sheehan & Messinis, 2003) A critical issue for such an assessment is whether there are obstacles to the evolution and upgrading of innovation systems, whether resource allocations, competencies and attitudes etc remain locked into patterns that are no longer productive. A recent comparative international survey found, as many similar surveys have found over the past 20 years, that, despite persuasive evidence to the contrary, Australian firms consider that their innovation-related performance is world class. 52 There is some evidence that public sector research organisations remain focussed on traditional fields of science, while the business sector is focussing on engineering and software, limiting effective 49 Where national economic, environmental, social (etc) performance is unsustainable and cannot meet the objectives of the society, change in the innovation system is one mechanism for achieving change in the wider economic, social and environmental system. 50 Bean, C. The Australian Economic Miracle : A view from the North. Australian Economy in the 1990s. Reserve Bank of Australia 2000 Conference. P Over the past decade users have captured the greater share of the benefits from ICT innovation. 52 Brown, D. The Innovative Firm. Arthur D. Little ISRN 6 th Annual Conference Working Paper 12

13 interaction. Bourke, drawing on data for the period, has shown that industry participation in scientific papers (and hence presumably also in research collaboration) is particularly low in Australia, about 2% compared with 8% in the UK and 9% in the US. 53 In relation to the third of these dimensions, one general area that appears to be a continuing systemic weakness is that of the exploration of new industry development through new firm formation, either as start-ups or spin-offs from existing firms. The relatively low levels of venture capital, particularly of early stage finance, appear to be a continuing problem. In relation to the third dimensions of assessment, it is vital to recognise that, despite the recent performance, the Australian economy remains vulnerable 54. Our analysis of trade, R&D and US patent data indicates that Australia has increased its relative specialisation in low-tech activities. There may well be a case for a more systemic and sustained approach to upgrading the accumulation of skills and knowledge and more generally ensuring that Australia has an innovation system able to contribute substantially to the development of the economy rather than simply respond to short term market signals. In this regard it is worth quoting at some length a recent discussion on the assumptions underlying the current policy settings: The new growth theories point out that the growth-enhancing effect of trade is an aggregate effect; we expect it to hold on average. but not in every case. In particular, trade can reduce growth for countries that have comparative advantage in industries with low-growth potential. Lower growth does not, however, necessarily imply lower economic welfare. Specialisation through trade may move the terms of trade in favour of the low-tech country which is enabled to import cheaper high-tech goods.trade is not, however, necessarily welfare enhancing in the absence of competitive markets. If there are substantial market failures in the accumulation of knowledge and skills and new goods, then trade is a double-edged sword. On the one hand, trade acts as a conduit for new ideas, stimulating growth and enhancing welfare. On the other hand, trade liberalisation and consequent specialisation in low-tech activities may relegate a country that is historically disadvantaged in the accumulation of skills and knowledge to fall further and further behind. The pessimistic view of trade liberalisation for Australia is that it might lead us to inefficient specialisation in natural resource based activities with few incentives for enhancing skills and knowledge. For example, the current recovery in the world economy is already having the effect of improving short-term prospects for the terms of trade and raising the real exchange rate. It is possible that such movements may squeeze out the recent expansion in exports of high valueadded manufacturing and lower our prospects for long-run growth and welfare by compounding failures to develop our skill and knowledge base. These are, however, second-best welfare arguments. It is not obvious that we should be using trade policy to rectify failures in the markets for the development of skill and knowledge and new goods. Rather, if we 53 The Allen Consulting Group. Systemic Mismatches in the National Innovation System and Bourke, P. (2000) Relative Strengths in Australian Basic Science. A Summary Bibliometric Map. Contributed Paper. National Innovation Summit. Bourke claims that in the US industry participated in one in four engineering and technology compared to one in 20 in Australia. 54 Like Sheehan and Messinis, Bean draws attention to the rising level of private and external Australian debt. Bean, C. The Australian Economic Miracle : A view from the North. Australian Economy in the 1990s. RBA 2000 Conference. ISRN 6 th Annual Conference Working Paper 13

14 address these problems directly, both the new theory and the econometric evidence suggest that trade liberalisation is likely to enhance both growth and welfare.[emphasis added] Australian Business R&D Regional Diversity and Changing Knowledge Base In , business R&D expenditure (BERD) was $4.8 billion. BERD is almost equally distributed between manufacturing and service industries (each accounting for about 45% of BERD), while mining accounts for about 10%. In , about 55% of all Australia s business R&D expenditure was undertaken within seven industry groups three service industries, three manufacturing industries, and mining (Chart 5). Australia s business R&D at the national level major fields of technological skill. The rising significance of ICT in Australian R&D is evident when we look at the fields of research. 56 In terms of the broad field of research, engineering and ICT account for over 80% of BERD. But within engineering the largest field of research is in communication technologies. Overall, business R&D expenditure directly focussing on R&D in information, computing and communication sciences, communications technologies, and computer hardware together comprise 38% of all BERD (Chart 6). Furthermore, half of the top eight fields of research (representing 65% of Australia s total BERD) are in ICT. Manufacturing and automotive engineering then follow, showing the traditional areas of the R&D skills base. Medical and health research skills are also significant, as is resources engineering a field strongly associated with the mining industry. Australian R&D expenditure at the level of States and Territories In comparison with leading economies, Australia s gross expenditure on R&D, at 1.53% of gross domestic product (GDP), is relatively modest. Corresponding data at the State and Territory level, however, varies considerably 57. Chart 7, comparing the trend in Australia s BERD as a proportion of GDP with corresponding trends (BERD as a proportion of Gross State Product (GSP)) for each state and territory, shows that Victoria has consistently had the highest BERD/GSP ratio (BERD intensity) of the states and territories. BERD in Queensland increased sharply in , but then followed the national trend. After falling in , SA s intensity has increased overall SA is the only state to have surpassed the peak level of BERD/GSP that most states and territories 55 Steve Dowrick Openness and Growth RBA 1994 Conference International Integration of the Australian Economy 56 The ABS requires firms to indicate the distribution of their R&D across both core business activity and field of research. 57 Not all of this variation in gross R&D expenditure as %GSP (Gross State Product) is due to differences in levels of business R&D expenditure Commonwealth and Higher Education R&D activity contributes to inter-state variation in R&D intensity. ISRN 6 th Annual Conference Working Paper 14

15 reached in the period to While R&D intensity declined after 1995 in most states the decline was particularly market in WA. Chart 5 BERD ( ) - Industry of Core Business Activity Other n.e.c. Scientific research Property and business services Finance and insurance Wholesale and retail trade Other Manufacturing Industrial Machinery and Equipment Electronic & Electrical Equipment etc Professional and Scientific Equipment Motor Vehicle etc Metal Products Non-Metallic Mineral Products Petroleum, Coal, Chemicals etc Printing, Publishing & Recorded Media Wood and Paper Products TCF Food etc Manufacturing Mining (Including services to mining) A$Million Chart 6: Australia Major R&D fields of technological skills (RF), (%GDP) ISRN 6 th Annual Conference Working Paper 15

16 Chart 7: Change in National and State-level BERD (as %GDP or % GSP) Chart 8 summarises a range of data on R&D performance and on patenting in Australia by organisations based the major states. It indicates the important sectors in each state in terms of R&D investment, R&D increase over the 1990s, the major fields of research, and relative patenting activity. On the basis of these patterns the characteristic strengths of each state are identified. In the following analysis of R&D patterns and trends at the state level we will discuss two States: New South Wales and Queensland 58. New South Wales NSW s gross expenditure on R&D (GERD) as a proportion of GSP ranks below the national GERD intensity average of 1.53%. This level results primarily from lower levels of R&D expenditure in Commonwealth agencies and in universities. However, NSW has Australia s second highest intensity of business R&D expenditure (0.71% of GSP), though it is significantly below the corresponding level in Victoria (0.98%). Chart 9 shows the major R&D performing industries in NSW: computer services; electronic equipment; finance and insurance; metal products; photographic and scientific equipment, and food processing. NSW s R&D skills base is dominated by computer software where it has the strongest research capability, both in scale and in R&D intensity. Communications technologies are also relatively strong being at approximately the same scale as in Victoria, but behind that State in R&D intensity in that field of research. 58 For a more detailed discussion of State and Territory patterns see Working Paper 3: Regional Aspects of Australia s R&D Activities ISRN 6 th Annual Conference Working Paper 16

17 Chart 8: Patterns of R&D Activity and Strength at State Level 59. New South Wales Victoria Queensland South Australia Western Australia Major R&D Performing Sectors Sectors with strong Growth in R&D Major Fields of Research Patent Strength & growth Apparent Overall Strengths ICT services; ICT equipment; finance and insurance; metal products; photographic and scientific equipment, and food processing. ICT services; Photographic & scientific equipment. Computer Software; Communication Technology; Medical; Electronics Processed food, instruments. ICT services & equipment; Instruments & devices ICT services; ICT equipment Chemicals; Auto. Instruments and parts; Medical & pharmaceuticals; Finance & insurance. ICT services; ICT equipment; finance & insurance Automotive, mechanical & industrial engineering; Communication tech.; Software; other ICT. Consumer goods & equipment; Mechanical engineering; Chemistry & life sciences Communication services & equipment; Automotive engineering, Biotech. Mining; ICT services; Metal products; Petroleum, coal and chemicals. ICT services Software, Metallurgy; Resource engineering; Auto., mech. & Industrial engineering. Civil engineering Mechanical engineering. consumer goods & equipment Resources, Software Petroleum, coal & chem.; Machinery & Equipment wholesaling; ICT equipment; Auto engineering; photographic & scientific equipment; ICT services. Machinery & Equipment ; Wholesaling; ICT services; ICT equipment ICT; Manufacturing engineering; Automotive engineering; Software Process engineering; instruments; Mechanical engineering Instruments; Auto. Engineering Metal ore mining; Property & business. Services; ICT equipment; ICT services. Metal ore mining Resource engineering; communication technology.; chemical engineering; metallurgy None Resources Overall, NSW is growing above the Australian average in most patent areas (Chart 11). Patent data points to relative strengths in Electronic equipment, instruments and processed food. These areas show strength in fine measurement and control of devices. 59 Based on the analyses in Working Paper 3: Regional Aspects of Australia s R&D ISRN 6 th Annual Conference Working Paper 17

18 Chart 11: NSW Major R&D industries (ANZSIC), to (%GSP) Chart 12: NSW business Major R&D technological skills base (RF), (%GDP) ISRN 6 th Annual Conference Working Paper 18

19 Chart 13: NSW Broad patent R&D fields, by relative emphasis and growth 1.30 Miscellaneous Electricity-electronics & IT Emphasis 1.20 Agriculture and food Instruments Civil eng, building, mining Consumer goods and equipment 0.90 Mechanical engineering Chemicals and life sciences Process engineering Electricity-electronics & IT Instruments Chemicals and life sciences Agriculture & food Process engineering Mechanical engineering Civil eng, building, mining Miscellaneous Consumer goods and equipment Queensland 0.60 Relative growth Queensland s R&D intensity (1.23% in ) is lower than other states. Queensland s BERD/GSP peaked at 0.57% in and stood at 0.47% in The mining industry has consistently been Queensland s largest R&D spender, but has recently been challenged by computer services (Chart 12). Chart 12: Qld Major R&D industries (ANZSIC), to (%GSP) Queensland s leading business research fields in were a close reflection of the two industries spending most on R&D, computer software was the leading research field followed by metallurgy and resources engineering (Chart 13). Queensland has a strong and growing emphasis on patenting in civil engineering and mining. There is also some emphasis and relative growth in the areas of mechanical engineering and consumer goods (Chart 14). ISRN 6 th Annual Conference Working Paper 19

20 Chart 13: Qld business Major R&D technological skills base (RF), (%GDP) Chart 14: Qld Broad patent R&D fields, by relative emphasis and growth Civil eng, building, mining Emphasis Consumer goods and equipment Mechanical engineering Agriculture and food Miscellaneous Electricity-electronics & IT Instruments Chemicals and life sciences Agriculture & food Process engineering Mechanical engineering 0.80 Process engineering Civil eng, building, mining Instruments Electricity-electronics & IT Chemicals and life sciences Consumer goods and equipment Miscellaneous 0.50 Relative growth ISRN 6 th Annual Conference Working Paper 20

21 Assessing BERD Performance in Australian States Taken overall, this analysis illustrates several points: Service sectors account for almost 50% of Australian BERD, and through the 1990s their share of BERD grew more rapidly than manufacturing, mining or agriculture. R&D in some manufacturing sectors (eg metal products in NSW and WA) declined sharply in the late 1990s. The role of ICT as a sector of industry (largely in services) and as a field of research is highly significant and pervasive. The top two fields of research in business are ICT and overall 65% of BERD is allocated to ICT-related research. In Computer services was the sector with the greatest level of investment in R&D and Communication services was the fifth largest R&D spending sector. The computer services industry showed the fastest (and by far the most consistent) rate of growth in R&D expenditure over the 1990s in NSW Queensland and SA. The second, third, fourth and sixth most important fields of research in industry in Australia in were all in engineering: communication engineering, manufacturing engineering, automotive engineering, and resource engineering. R&D in software and engineering account for the majority of R&D industry but a minor share of R&D in the public sector. The possibility of a mismatch in research allocation and in human resource development needs to be assessed. Mining was the second most important sector in terms of BERD expenditure in , the most important sector in Queensland and the NT, and dominated all R&D effort in WA. Patterns of R&D activity and directions of change in that activity vary significantly among the states. 5. Australian Innovation: Patterns of Specialisation and Evolution. Much of the comparative analysis of national innovation systems has focused on the research system (Nelson, 1993). This focus characterises a nation s technological specialization by assessing the level and direction of innovative effort. Since such approaches focus on only some dimensions of the innovation system they provide a narrow view on the overall innovation system (Lundvall, 1992). ISRN 6 th Annual Conference Working Paper 21

22 The following analysis of the recent (20 year) evolution of Australia s technological specialization uses indicators based on R&D, patenting and scientific publications 60. We compare Australia s performance with that of several OECD countries and in particular with Canada and Finland, as these two countries share a similarly long history of specialisation in natural resource-based sectors. The Evolution of the Pattern of the Innovative Effort as indicated by R&D Expenditure Australia s R&D expenditure pattern is characterised as low in GERD and BERD, but high in the government share of total R&D expenditure (Gregory, 1993; Australia Science & Technology at Glance, DEST, 2002). Chart 15 shows total R&D intensity levels (GERD/GDP) over for Australia, Canada, Finland and the overall OECD average. Total R&D intensity levels in Australia remained below the OECD average. In the early 1980s Canada, Finland and Australia had roughly comparable R&D intensities. While levels in Canada and Australia have remained similar, by the late 1990s Finland s R&D intensity was more than double that of Australia. Chart 15 R&D Intensity : Australia, Canada, Finland & the OECD Finland Canada Australia OECD Source: Ausis Patterns of Sectoral Distribution of Business Funding for R&D In all three countries over period 1974 to 1997 R&D has grown particularly strongly in some sectors, (eg Electronics in Finland, Communications Services and Pharmaceuticals in Canada, and Commercial & Engineering Services in Australia). But only in Australia has it grown in all sectors. In both Canada and Finland some significant sectors showed very slow growth or a decline in real 60 These indicators have significant limitations and need to be interpreted with caution. See Ausis (2003) Working Paper 4. Assessing ISRN 6 th Annual Conference Working Paper 22